Frequently Asked Questions

No. Open the reader, paste some text, and start reading. A free account adds saved documents, flashcards, and progress stats.

Mandarin Chinese, Cantonese, Vietnamese, and Thai. Each one is segmented and colored by tone using language-specific dictionaries and word lists.

Peer-reviewed research says yes. A 2017 randomized study by Godfroid, Lin & Ryu (Language Learning, 67(4)) found that pairing each tone with a consistent color measurably improved tone perception in beginners compared to audio-only training.

Each tone in a language is assigned a consistent color — for example, Mandarin tone 1 (high level) is one color and tone 4 (falling) is another. Every character you read is colored by its tone, so your eye builds an association between color and sound over time without any extra effort. Research supports this approach: a 2017 randomized study by Godfroid, Lin & Ryu (Language Learning, 67(4)) found that multimodal visual cuing including color significantly improved tone perception in novice learners.

Chinese (and Thai) text has no spaces between words. ToneReader uses a word-list-based segmenter to find word boundaries, so clicking on 公园 shows the compound "park" definition rather than the individual characters 公 and 园 separately.

ToneReader uses FSRS-5 (Free Spaced Repetition Scheduler), the most accurate open-source SRS algorithm available. Unlike the older SM-2 algorithm used by early Anki versions, FSRS models two separate properties for each card: stability (how long you can go before forgetting) and difficulty (how hard the card is for you personally). When you grade a card Again / Hard / Good / Easy, FSRS updates both values and schedules the next review to target 90% retention — meaning you see each card just before you'd forget it. The result is fewer reviews for the same level of retention compared to SM-2.

Texts up to 20,000 characters, 10 saved documents, and 50 flashcards. It does not expire, and no card is required to start.

Audio is a Pro feature. ToneReader sends the requested word or phrase to Microsoft Azure Cognitive Services (neural text-to-speech). The resulting audio is cached on Cloudflare R2 so repeated playback is instant and does not re-call the API.

Yes. Go to Flashcards → Export. You can download all cards (or a single language) as a CSV spreadsheet or as an Anki-compatible tab-separated text file (.txt) importable directly into Anki.

PDF, EPUB, DOCX, RTF, plain text, Markdown, CSV, SRT/VTT subtitles, HTML, and XML. Rich formats (PDF, EPUB, DOCX) are extracted on our servers; plain text formats are processed entirely in your browser.

Not yet. The web app is designed to work well on mobile browsers. A dedicated iOS/Android app is on the roadmap.

Visit our support page and select "Account deletion request" as the topic. All your documents, flashcards, and personal data will be permanently erased within 30 days.

Use the contact form on our support page. We aim to respond within 2 business days.